LEVERAGING THE SNBENCH INFRASTRUCTURE

1
LEVERAGING THE SNBENCHINFRASTRUCTURE

• MARIANNE PROCOPIO
  SMITHA PATIL
• rbtr1_at_cs.bu.edu
  smithap_at_cs.bu.edu

2
Objective

• To count the number people in a region.
• Approach To detect the change in temperature
  and correlate that with the number of people,
  using image processing techniques.

3
Milestones

• Java code and initial testing completed March
  19th
  
• Java package completed and the technical report
  completed April 25th
• Presentation completed. Source code, technical
  report and website submitted May 1st

4
Organization

• Program Design
• Smitha and Marianne
• Website Set Up and Updates
• Smitha and Marianne
• Source Code and Inline Documentation
• Marianne
• Technical Report and Testing
• Smitha

5
Steps
Temperature change
Image Acquisition

Image Subtraction
Threshold
Blob
Counter
6
4
7
ALGORITHMS

• Background Subtraction -gt Binary Image

-

• Connected Components -gt Blob Count
• Additional Opcodes lt, gt, toDouble, toGrayscale

8
RESULTS

• Background New Data Final Result

9
LIMITATIONS

• Camera Movement
• Incorrect Threshold
• -For Binary Image, Blob Count and Temperature
  Sensor
• Lighting Conditions
• Blob Miscounts
• Additional Image Processing Problems

10
Comparative Analysis

• Difference From Previous Applications
• Extend previous STEP Programs
• Provide Analysis of Data
• New and Modified Opcodes

11
FUTURE WORK

• Address Limitations
• Multiple Background Images
• Adaptive Thresholds
• Use Additional Image Processing Algorithms
• Correlate Temperatures with the Number of People
  Detected.