Presented by Yi-Ta Wu April. 23, 2004 1 Agenda The Surveillance - - PowerPoint PPT Presentation

presented by yi ta wu
SMART_READER_LITE
LIVE PREVIEW

Presented by Yi-Ta Wu April. 23, 2004 1 Agenda The Surveillance - - PowerPoint PPT Presentation

New Jersey Institute of Technology Presented by Yi-Ta Wu April. 23, 2004 1 Agenda The Surveillance System Auto-Alarm Based Surveillance System Two Problems of Surveillance System Wireless Sensor Based Surveillance System


slide-1
SLIDE 1

1

New Jersey Institute of Technology

  • April. 23, 2004

Presented by Yi-Ta Wu

slide-2
SLIDE 2

2

The Surveillance System Auto-Alarm Based Surveillance System Two Problems of Surveillance System Wireless Sensor Based Surveillance System Conclusions

Agenda

slide-3
SLIDE 3

3

The Surveillance System

  • The surveillance system is a series of monitoring devices

designed to check on environmental conditions.

  • It is widely used nowadays to help a guard with consecutive

sensing information.

slide-4
SLIDE 4

4

Problem of Current Surveillance System

Exhaustion by long-term observation

slide-5
SLIDE 5

5

Auto-Alarm Based Surveillance System

Auto-Alarm

slide-6
SLIDE 6

6

What is the Proper Resolution ?

Goal: Perform the face recognition Problem: It is difficult to determine the proper resolution.

slide-7
SLIDE 7

7

First image frame Second image frame The difference of above two images

What is the Target Object?

slide-8
SLIDE 8

8

Room 4 Room 5 Room 6 Room 1 Room 2 Room 3 Camera

Sensor 6 Sensor 1 Sensor 2 Sensor 3 Sensor 4 Sensor 5

Wireless Sensor Based Surveillance System

,

∑ ∑ ∑ ∑

= =

m y m m x m

i i i i i i

Y X

slide-9
SLIDE 9

9

Object Classification, Human/Non-Human

) (

1 +

− = ∆

t t

I I abs ) (

1 ↑

− =

+ t t

I I abs U ) (

1 ←

− =

+ t t

I I abs L ) (

1 →

− =

+ t t

I I abs R ) (

1 ↓

− =

+ t t

I I abs D

  • P. Viola, M. Jones, and D. Snow, “Detecting Pedestrians Using Patterns of

Motion and Appearance,” Computer Vision, 2003. Proceedings. Ninth IEEE International Conference on, pp. 734-741, Nice, France, Oct. 2003.

slide-10
SLIDE 10

10

(a) (b) (c) (d)

The Method for Efficiently Extracting Candidates

Program Demo

slide-11
SLIDE 11

11

Sets of Candidate for Face Recognition

slide-12
SLIDE 12

12 SVM Classifier Discrete Wavelet Transformation (28 * 23) 2D-LDA Reduction Yi-Ta Wu

The Flowchart of the experimental procedure

Original Image (112 * 92) Histogram Equalization

Sets of Candidate for Face Recognition

slide-13
SLIDE 13

13

Yi-Ta Wu

(a) (b) (c) (d)

Human

Experimental Results

slide-14
SLIDE 14

14

Conclusions

The traditional surveillance systems are performed either periodically

  • r manually rotating the cameras, which may have the difficulties of

determining the proper object resolution, and the automatically detecting the moving object when the camera is not fixed.

Wireless Sensor based surveillance system:

  • The Wireless Sensors are utilized as the first layer guards to detect the

coordinates of the unauthorized invasions

  • After identify the location of the invasions, the system will arrange

cameras to catch the image features for object classification and recognition based on the hierarchical approach.

slide-15
SLIDE 15

15

Question