SMART CAMERA SURVEILLANCE SYSTEM AND VIDEO ON DEMAND

1
SMART CAMERA SURVEILLANCE SYSTEM AND VIDEO ON
DEMAND

• UNDER Prof. Anupam Agarwal E Manuel (ADRIN)

2
Main Module

• Real Time Change Detection
• Real Time Video Transmission
• Serve Video on demand

3
REAL TIME CHANGE DETECTION

• A stand alone computer which will capture real
  time stream of images through web cam
• A background program continuously running to
  capture the images.
• Capture any change detection
• Categorize change detection
• Transmit video to Web Server

4
REAL TIME TRANSMISSION

• Real time stream of images are moved from source
  machine to web server
• Web server is connected to different computer in
  LAN and WAN
• Transmission of video stream in real time using
  JMF (Java Media Framework) using RTP (Real Time
  Protocol) to Web server and to any computer
  connected in LAN there after

5
SERVE VIDEO ON DEMAND

• Now any computer connected to the web server can
  watch or see the video frame
• A apache web server is set up in order to
  transmit to images which can be browsed by
  anyone from anywhere in the world

6
PROJECT OBJECTIVE

• Basically project main aim is to identify change
  occurred when any human come in the scope of
  scene.
• Once the human is detected it is segmented from
  the image.
• And the human motion is categorized in different
  category. Walking (L, R) (R, L), Sitting, Moving
  right hand, Moving left hand.

7
PROJECT OBJECTIVE

• Now once the change occurred and identified that
  change is due to human than the video stream is
  transmitted to Web server.
• From web server any one sitting in the remote
  place can access the recent images.

8
CHANGE DETECTION

• Change Detection comprises of online detecting of
  any change taking place.
• Change Detection must be luminous Invariant
  background Invariant, false change Detection
  Invariance
• Change detection is done in Static web camera if
  someone moves web cam than it can lead to false
  change detection

9
CHANGE DETECTION

• There are various problem which are faced for
  motion detection, and various algorithm are
  proposed for there solution.1234.
• Since the main task is Human Motion Detection lot
  of work has been done for that.56
• In my project I have combined various algorithms
  of both change detection and human motion
  detection to make my project more robust.

10
HUMAN MOTION DETECTION

• In my project, various algorithm in 1 is
  implemented for change detection. Once change is
  detected than segmented image is moved to second
  part that is Human Motion Detection.
• Further than various algorithm in 56 will be
  used to categorize the motion of human and
  identify that change is occurred due to human
  only or not

11
HUMAN MOTION DETECTION

• Basically both change detection and human motion
  detection is used for my project to increase
  efficiency, accuracy, and will also decrease time
  to execution as instead of detecting human motion
  in each frame first changed detection is
  performed and if change occurred than it move to
  next module which look for possible human in the
  image.

12
JMF

• Java Media Framework is the API of Java which use
  RTP to transmit online video stream and audio
  stream.
• JMF provide data sources, in other word JMF have
  API which are used to construct the function
  performed by common data sources78

13
JMF
14
HTML and WEB SERVER

• Further in order to transmit online video stream
  and for any system to access online change
  detected, applet using html is developed.
• Web server is the server where apache web server
  is kept from where any client can access the
  desired web page.

15
Work Done

• Online capture of image and video stream with the
  help of web cam.
• Online change detection irrespective of back
  ground, light invariance.
• Server Side of JMF file to transmit the video
  stream.

16
PLAN OF WORK

• Starting date- 25 Jan
• Study of documentation-25 Feb.
• Change detection and motion detection- 31 march
• Video on demand using JMF- 30 April
• Applet and testing- 10 May

17
REFERENCES

• 1 Dynamic Vision - R. K Jain R. Kasturi
• 2 IlluminationInvariant Change Detection -
  Daniel Toth, Til Aach, and Volker Metzler
• 3 Adaptive Change Detection for Real-Time
  Surveillance Applications -Stefan Huwer,
  Heinrich Niemann
• 4 Thresholding for Change Detection-Paul L

18
REFERENCES

• 5Towards Detection of Human Motion-Yang Songy,
  Xiaolin Fengy and Petra Peronay
• 6A Computational Model for Motion Detection and
  Direction Discrimination in Humans-Yang Song
  and Pietro Peronatj
• 7 java.sun.com/products/java-media/jmf/2.1.1/gui
  de/RTPRealTime.html
• 8www.java.sun.com/products/java-media/2D/index.h
  tml

19
THANK YOU