Much work has been done over the years to find ways of generating a virtual image of an object in 3D. Most stereoscopic methods which currently exist, attempt to match every point from an image taken from a left image, to every point on an image taken on a right image. Not only is this very slow unless a great amount of computational power is used, but also the result is worthless when attempting to describe the object to a machine. At best, the distance to an obstacle can be calculated. The methods I have been experimenting with are able to generate vectors of the left and right camera images to represent the 2D projection, and then attempt to reconstruct the 3D object by matching together the vectors. If this works correctly, it should allow the object to be shown in 3D space by a series of vector lines. These could be interpreted by a higher-level system to recognize the object, determine orientation and position of the object in 3D space, and make sense of simple scenes.

Grade Grouping: 9
Team Size: 1
Subject Area: Computer Science
Project Type: Descriptive
Project Level: Advanced
Project Format: Traditional

Software Tools:

• Website:
  • Macromedia Dreamweaver MX
  • Macromedia Flash MX
  • Macromedia Fireworks MX
  • Adobe Photoshop 7
  • Adobe Acrobat 5
  • Microsoft Office 2003
• Project:
  • Visual Basic 3.0
  • QBASIC 4.5
  • MPLAB IDE 6.40: MPASM Assembler
  • Borland C++

Special Skills:

• Website:
  • HTML
  • Flash ActionScript
  • JavaScript
• Project:
  • Knowledge of Visual Basic, QBASIC, MPASM Assembler, and C++
  • Circuit Design
  • Printed Circuit Board Layout
  • Soldering
  • Circuit Troubleshooting
  • Software Debugging
  • Microcontroller Programming
  • Algorithm Design and Testing
  • Love of Coffee :)

Hardware Tools:

• Website:
  • PC with Windows XP
  • PC with SuSE Linux 9.0
• Project:
  • PC with Windows 2000
  • PC with Windows for Workgroups 3.11
  • PicStart Plus Microcontroller Programmer
  • MPLAB ICD2 Microcontroller Debugger
  • 30MHz Oscilloscope
  • Multimeters
  • Logic Probe
  • Power Supply

Source of Idea:
I originally thought of the idea for this project about two years ago. I had created 3 projects at the time, all related to electronics. I had recently experimented with PIC microcontrollers and stepper motors, so I wanted, in some way, to include those. I was also interested in doing something related to artificial intelligence or robotics. As I did some research on current methods and algorthms, I narrowed this topic down to specifically robot vision. A goal I have is to expand on this project to allow for object recognition, sometime in the near future.

Awards Recieved:

• Calgary Youth Science Fair (March 2004)
  • Gold Medal
  • Noel Bourget (Auto-trol) Award
  • Nortel Networks Technology and Innovation Award (for seconday project)
  • Government of Alberta Travel Scholarship
  • Top Intermediate Project Award
• Canada Wide Science Fair (St. Johns, Newfoundland, May 2004)
  • Place on Team Calgary Awarded (fair has not yet taken place)

Website: http://www.alumni.ca/~stag4m0 - posted on Virtual Science Fair website at http://www.virtualsciencefair.com.