0% found this document useful (0 votes)
68 views2 pages3GSS Intelligent Tracking System Using CAMSHIFT Algorithm: Rupali R Sonawane Prof.S.O.Dahad
This document summarizes an intelligent tracking system that uses the CAMSHIFT algorithm to track people of interest in real-time video from camera modules. It aims to avoid human errors in monitoring. The system performs moving object detection using background subtraction and then applies the CAMSHIFT algorithm for tracking. The CAMSHIFT algorithm is described as being simple and efficient for real-time tracking compared to other algorithms like mean shift that have higher computational complexity. The proposed system architecture incorporates multiple camera modules connected to a central display for security monitoring personnel.
Uploaded by
Editor IJRITCCCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
68 views2 pages3GSS Intelligent Tracking System Using CAMSHIFT Algorithm: Rupali R Sonawane Prof.S.O.Dahad
This document summarizes an intelligent tracking system that uses the CAMSHIFT algorithm to track people of interest in real-time video from camera modules. It aims to avoid human errors in monitoring. The system performs moving object detection using background subtraction and then applies the CAMSHIFT algorithm for tracking. The CAMSHIFT algorithm is described as being simple and efficient for real-time tracking compared to other algorithms like mean shift that have higher computational complexity. The proposed system architecture incorporates multiple camera modules connected to a central display for security monitoring personnel.
Uploaded by
Editor IJRITCCCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
/ 2
International Journal on Recent and Innovation Trends in Computing and Communication
Volume: 4 Issue: 3
ISSN: 2321-8169
184 -185
______________________________________________________________________________________
3GSS Intelligent Tracking System using CAMSHIFT Algorithm
Rupali R Sonawane
Dept of E&TC
Government college of Engg.
Jalgaon,Maharashtra
rups4892@gmail.com
Prof.S.O.Dahad
Dept of E&TC
Government college of Engg.
Jalgaon,Maharashtra
dahad.sanjay@gmail.com
AbstractThe aim of this system is to track a person of interest in real time based on scenes obtained from camera modules. This system avoids
potential human errors as we are using automatic monitoring. It is a histogram based real time tracking system. Moving objects are detected
using background subtraction.Moving object detection, recognition and tracking are basic steps in processing video frames.
Keywords-tracking,surveillance,CAMshift
__________________________________________________*****_________________________________________________
I.
INTRODUCTION
Public security has become very important issue due to
the increased terrorist attacks, burglaries etc. Electronic
surveillance forms an integral part of todays security
systems. CCTV modules are sold in huge numbers across
the globe. Primitive surveillance systems [1] involved
cameras. All these cameras would feed the live video to a
dedicated monitor and in a control room many such
monitors would be stacked one upon the other. These
systems can be found in departmental stores, museum to
high security areas like railway stations, airports and banks.
Figure 1 . Applications of camera surveillance systems
The main objective of this project is to develop a system
which reliably detects the persons presence or absence in a
locality.The first part of the paper provides information
about thealgorithm selection and camera selection.
II.
LITERATURE REVIEW
In [2] a method to track any object which is differsfrom
thebackground
was
proposed.
Multiple
objects
getconsistently tagged and hence it is a problem of selecting
an object of interest.In [3] a method of tracking an aerial
object was discussed. The given image is divided into
blocks and these blocks are adaptively changing their size
according to template size. In [4,5] authors have discussed
motion segmentation computation and also optical flow. It
creates a difficulty when object is detected at start. In [6] a
real time tracking and adaptive framework was proposed. In
[7] partial filtering was discussed. In [8] author discussed
mean shift algorithm which is suitable for multiple objects
tracking. In [9] it showed that these algorithms are
computationally complex. In [10]CAMshift algorithm was
discussed, which issimple and efficient algorithm for
tracking.In [11] location of the object could be tracked from
probability distribution of an image. In [12] the author
proposed an algorithm which processes all the pixels in a
given frame so it increases computations.
III.
VIDEO PROCESSING ELEMENTS
Methods for noise reduction
Noise is unwanted signal in a video capture which degrades
the video quality. To reduce the effects of noise in the video
capture, the following methods will be useful:1. Using low pass filtering method: Noise may create sharp
variations in an image. Some parts of image get changed due
to noise. So low pass filtering can beuseful.
2. Using averaging method: In this method of noise
reduction, 3 or 4 consecutive frames of videos are taken and
added pixelby pixel and then averaged. This will reduce the
random noise.
Color space conversion
Camera modules provide the video in any color format.
System converts the video from one color format to another.
In image pipeline YUV color space is typically used. Taking
human perception into account it encodes the color format.
RGB color space is also simple to understand and it has all
possible colors. It is made up of three colors: red color,
green color and blue color.
IV.
SYSTEM ARCHITECTURE
The application scenario would include an enclosed building
where people enter one at a time. The building is equipped
with many interconnected camera modules. Each of the
entrances is to be provided with a camera system such that
each entrant is registered by taking his high resolution
photograph. These images of people are stored in a central
storage. Whenever a query for a person is requested in terms
of photograph, the supplied image is compared with the
ones in the database. After verifying the persons presence,
the feature related data of the subject is communicated to all
the camera modules for tracking. The status of the tracked
184
IJRITCC | March 2016, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication
Volume: 4 Issue: 3
ISSN: 2321-8169
184 -185
______________________________________________________________________________________
person is conveyed to the security official via a display
terminal as shown in Figure1.
B. Mean-shift algorithm
Mean shift algorithm is used for finding the maxima of a
density function.The mean shift vector always points
towards the direction where there is maximum density or
where the gradient of density function is zero.
C. Adaptive mean shift algorithm
The performance of mean shift tracking reduces in real time
tracking as it uses a fixed window size. As the name
indicates adaptive mean-shift algorithm uses adaptive
window sizes. The window size changes depending on the
features matching. This algorithm internally uses mean shift
algorithm.
VI.
Figure 1 . Block Diagram
Camera modules
The camera modules arelocated at various locations in an
enclosed place or a building. As a high resolution image
would require larger computations while a low resolution
image would degrade the video quality. Hence, an optimum
resolution of 640x480 is a good trade off. The camera must
present the data in an easily readableRGB for further
processing on the host machine.
V.
PROPOSED TRACKING ALGORITHM
The objective of this system is to track the moving person
across multiple cameras. Mean shift algorithmis used for
tracking. Mean shift algorithm is less complex and requires
few computations to implement in real time. As the system
needs to be real time mean shift algorithm is suitable. The
limitation of mean shift algorithm is that it cannot detect
change in size and orientation of objects. This can be solved
by using adaptive mean shift algorithm. If the object is
occluded then mean-shift algorithm fails. This can also be
addressed by adding background subtraction into mean-shift
algorithm. For using mean-shift algorithm for tracking the
color information is required. This color information is
known as histogram.
A. Histogram
An image histogram is a graphical representation of the
color distribution in image. It plots the number of pixels for
each color intensity. By looking at the histogram for specific
image, a viewer will be able to judge the entire distribution
at a glance.
.
Figure 2 . Image and its histogram
CONCLUSION
In this paper, we have presented new intelligent system for
real time person tracking and described how the system can
be used to track down person in a locality. The system is
designed to receive the data fromcamera modules. We have
integrated mean-shift algorithm to achieve adaptive window
size.
REFERENCES
[1]
T.D. Raty. Survey on contemporary remote surveillance systems for
public safety. Systems, Man, and Cybernetics, Part C: Applications and
Reviews, IEEE Transactionson, 40(5):493515, 2010.
[2] Y. Wang, M. Casares, and S. Velipasalar, Cooperative object tracking
and event detection with wireless smart cameras, in Advanced Video
and Signal BasedSurveillance, 2009.AVSS 09. Sixth IEEE
International Conference on, 2009, pp. 394399.
[3] S. Sahani, G. Adhikari, and B. Das, A fast template matching
algorithm for aerialobject tracking, in Image Information Processing
(ICIIP), 2011 International Conferenceon, 2011, pp. 16.
[4] Jung Uk Cho, Seung-Hun Jin, Xuan Dai Pham, and Jae WookJeon.
Object trackingcircuit using particle filter with multiple features. In
SICE-ICASE, 2006. InternationalJoint Conference, pages 14311436,
2006.
[5] S. Denman, C. Fookes, and S. Sridharan.Improved simultaneous
computation of motion detection and optical flow for object tracking.
In Digital Image Computing:Techniques and Applicat ions, 2009.
DICTA 09., pages 175182, 2009.
[6] X. Chen, X. Li, H. Wu, and T. Qiu, Real-time object tracking via
CAMshift-basedrobust framework, in Informat ion Science and
Technology (ICIST), 2012 InternationalConference on, 2012, pp. 527
530.
[7] J. U. Cho, S.-H. Jin, X. D. Pham, and J. W. Jeon, Object tracking
circuit using particle filter with multiple features, in SICE-ICASE,
2006. International JointConference, 2006, pp. 1431 1436.
[8] B. Li, Z. yuanZeng, and Z. ru Wu, Multi-object tracking based on
improved mean-shift algorithm, in Image and Signal Processing,
2009. CISP 09. 2nd InternationalCongress on, 2009, pp. 15.
[9] J. U. Cho, S.-H. Jin, X. D. Pham, J. W. Jeon, J.-E.Byun, and H. Kang,
A real-timeobject tracking system using a part icle filter, in
Intelligent Robots and Systems, 2006IEEE/RSJ International
Conference on, 2006, pp. 28222827.
[10] G. Bradski, Real time face and object tracking as a component of a
perceptual userinterface, in Applications of Computer Vision,
1998.WACV 98.Proceedings.,FourthIEEE Workshop on, 1998, pp.
214219.
[11] X. Qin-lan, A real-time motion target detecting and tracking system
based on tms320dm642, in Intelligent Systems, 2009. GCIS 09. WRI
Global Congress on, vol. 3,2009, pp. 8387.
[12] X. Gang, C. Yong, C. Jiu-jun, and G. Fei, Automatic CAMshift
tracking algorithmbased on fuzzy inference background difference
combining with twice searching, in EHealthNetworking, Digital
Ecosystems and Technologies (EDT), 2010 InternationalConference
on, vol. 1, 2010,pp. 14.
185
IJRITCC | March 2016, Available @ http://www.ijritcc.org
_______________________________________________________________________________________