CIS 543 Presentation 3

1
CIS 543 Presentation 3

• Team 3
• Unmanned Blimp

2
Team 3

• Christopher Lowdon (TL)
• William Hitchcock
• Julie Monaco
• Mark Troutfetter

3
Overview

• Project Introduction
• Issues Log
• Effort Evaluation
• Project Design
• C0 Testing
• Project Demonstration
• Conclusions and Strategies

4
Project Introduction

• The purpose of the Unmanned Blimp project is to
  design a blimp that can follow a physical path
  without any human intervention.
  
• Capture image from Blimp perspective
• Process image of line on the floor
• Detect edges
• Calculate displacement of blimp
• Send signal for blimp movements

5
Issues Log Resolved

• Blimp stuck moving Forward in Automatic mode
• Debugged the code
• Found it wasnt comparing the two halves of the
  image correctly
  
• Behind Schedule
• Put in more work on the weekends
• Planned out next steps ahead of time

6
Issues Log Resolved

• Blimp will not support the camera
• Used Mylar to make a bigger blimp
• How to process the video feed
• Discussion with Dr. Day
• Used a histogram pixel sampling method
• Determines where the largest concentrations of
  black pixels are.

7
Issues Log Unresolved

• Process does not close when exiting the program
• Switch to Manual mode before exiting
• More debugging needs to be done
• If Camera is disconnected unrecognizable error
  is thrown at start of program
  
• Try to display No video device found error
  first
  
• Check to see if we can open the program first and
  then show error.
  

8
Issues Log Unresolved

• Blimp does not tightly follow a line
• More advanced line following algorithm needed
• Sample more or all of the pixels in the picture
• Fan motors are not large enough to move Blimp in
  timely manner
  
• Get larger fan motors
• Get a new Blimp

9
Effort Evaluation Time Logs

• Christopher Lowdon
• Total Hours 118.25 Hours
• Sample of Chriss Time Log

10
Effort Evaluation Time Logs

• William Hitchcock
• Total Hours 74.25 Hours
• Sample of Willies Time Log

11
Effort Evaluation Time Logs

• Julie Monaco
• Total Hours 72.5 Hours
• Sample of Julies Time Log

12
Effort Evaluation Time Logs

• Mark Troutfetter
• Total Hours 70.5 Hours
• Sample of Marks Time Log

13
Effort Evaluation
14
Effort Evaluation
15
Effort Evaluation
16
Effort Evaluation
17
Effort Evaluation
18
Code Breakdown

• Christopher Lowdon
• C0 coverage
• Blimp Command
• Mark Troutfetter
• Blimp Command
• Drive() Method
• Willie Hitchcock
• Integration of Code
• Manual Mode
• GUI
• Julie Monaco
• GUI
• About Box
• Histogram

19
Project Design

• Flow Charts
• Hardware/Software
• Object Model
• Sequence Diagrams
• Algorithm Discussion

20
Flow Chart
AUTOMATIC MODE
Capture Image from Blimp
Calculate Blimp Displacement
Send Movement signals to Blimp
Detect Edges
21
Flow Chart
MANUAL MODE
NO
0 DOWN 1 UP 2 Backward 4 Turn Left 6
Turn Right 7 Spin Left 8 Forward 9 Spin
Right

Capture Image from Blimp
Check for Keyboard Input
Send Movement Signals to Blimp
Display Image
YES
22
Hardware

• Bigger Blimp
• Mylar Material
• RC Controller
• SN74LS07N Hex Buffer
• Parallel Cable
• Camera
• Wireless Color Spy Camera
• Video Capture Card
• USB Live

23
Software Application

• Program stillCap.exe
• Video Capturing Software
• Capture images (uncompressed)
• Communication with RC Controller
• Via Parallel Cable
• Edge Detection
• Histogram Pixel Sampling Algorithm
• Blimp Displacement Calculation
• Compare concentrations of Black pixels (from
  Sampling Algorithm)
  

24
Software Application

• GUI
• Display Original Image
• Display Filtered Image
• Button to switch modes and display it
• Display Movement
• 2 Modes
• Automatic Mode
• Detects Edge from Filtered Image
• Determines position of Black Line
• Moves Blimp Accordingly
• Manual Mode
• Check for Keyboard Input
• Move Blimp Accordingly

25
Object Model
26
Sequence Diagram
Manual Mode
27
Sequence Diagram
Automatic Mode
28
Filtering Algorithm

• Filter Image into Gray Scale
• 3 elements/pixel in bitmap array (RBG)
• Average the 3 elements store average back into
  each of the 3 elements
  
• Histogram Pixel Sampling Method
• Create Histogram of all the shades of gray from
  0-255
  
• Use top 20 percent as the cut off for Black
  Pixels
  
• Set all pixels above cut off to 255 (White)
• Set all pixels below cut off to 0 (Black)

29
Edge Detection Algorithm

• Using Histogram Pixel Sampling Method
• Sample Right 1/3 and Left 1/3 and count number of
  Black pixels in each
  
• Comparison within Threshold
• Move ? Forward
• Comparison above or below threshold
• Move ? Right or Move ? Left

30
Testing

• Formal Test December 1, 2005
• Tested Blimp Functionality (Next Slide)
• Reviewed by Team Lead Chris
• I give the Unmanned Blimp Project two thumbs
  UP!
  
• C0 Testing December 5, 2005
• Passed almost all tests
• Included in documentation due Friday 12/9/05

31
Testing
32
Project Demonstration

• Manual Mode
• Starting screen
• Display live video feed
• Display filtered image
• Display directional commands transmitted to
  blimp
  
• Allow manual blimp control

33
Project Demonstration

• Automatic Mode
• Enter after the start button is pushed
• Continue live video feed
• Capture a single image and pass its pixel values
  into a buffer
  
• Pass buffer into filter making each value either
  white, gray or black.

34
Project Demonstration

• Automatic cont.
• Display filtered buffer to screen as an image
• Pass filtered buffer to an image processing
  function that will return a command to be sent to
  the blimp
  
• Send command to blimp
• Start automatic process over by capturing a new
  image from the video feed
  

35
Project Demonstration

• Once stop is pressed the program returns to
  manual mode and waits for user to start Automatic
  mode again.

36
Future Developments

• Determine Height of Blimp
• Line Width
• More Advanced line following
• Account for different turns
• Try to find line if goes off screen

37
Conclusions

• Gained team management skills
• Time analysis, efficiency, communication skills,
  project planning
  
• Learned Software Engineering techniques
• Gantt chart, testing, code integration, modeling,
  software security
  
• Project was a success!
• Reason Hard work, time, and good people skills.