Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Computer Vision Tracking

(27/05/13) Richard Radic
by

Richard Radic

on 30 May 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Computer Vision Tracking

Computer Vision Tracking By: Richard Radic Introduction Future Work Objective Implementation Conclusion Testing Matlab implementation Agenda
Introduction
Objective
Main components
Algorithm design
Implementation
Testing
Future work
Questions and Demonstration Questions? What is computer vision tracking? Method by which we automate machines to track moving objects using interfaced cameras. Where is computer vision tracking? Premises security
Factory production lines
Military applications How is Computer vision tracking implemented? Algorithm design implementation
Operational requirements
Platform specification Objective of the project:
'Track a Red Ball in real time with the use of an ARM based computer.' Raspberry Pi - ARM based computer
Web cam
Red Ball Initial investigation
Basic Image Processing
Still Image Manipulation
Matlab© Raspberry Pi implementation
Linux - Debian Familiarisation
Python - Basic 'Image Processing' programing Tracking algorithm -
Development, implementation and refinement Procedure cycle Testing:
Prototypes
Headings - Lighting, Efficiency, Accuracy etc. Project Components Image Processing Raspberry Pi Matlab Python Hardware Software Image Processing
Raspberry Pi Image Processing
Matlab imaging library
Extensive help library Python imaging library
Pygame
Numpy/ Scipy Image format
Image filtering
Object segmentation
Algorithm design 4G SD card for operating system
and internal memory
500MB RAM
LAN internet access
HDMI output
USB HUB
Keyboard, Mouse and Web cam Linux
Raspbian (Debian)
Geany IDE
Python Programming
Language Three Main components:
Locking Mode
Tracking Mode
Recovery Mode Object identification
Whole image scan
RGB thresholds
Conversion to Binary (black and white image)
Calculate Mean of detected region
(centroid) Uses obtained Centroid values
Calculates radius
New scan region (efficiency)
Error avoidance -
scan area boundaries Algorithm Design Tracking Mode Recovery Mode Locking mode: Object is lost from scan region
Extend scan region or reduce image resolution (efficiency)
Identify object - resume Tracking mode
Otherwise Implement Locking mode Main components Raspberry Pi implementation Project supervisor: Mr Gerard Egan Efficiency Time module
Implementing timer on all aspects of script
Implementing second scan loop - Tracking
'np.where' function
Image input dimensions (i.e. 320*240)
Image format Lighting conditions Light intensity meter (foot candles - luminous flux)
Light changes causes change in RGB thresholds
Types of light:
Direct sunlight - 100000 lux
Daylight - 1600 lux
Shade - 215 lux
Artificial light - 258 lux

Solution: Dynamic Thresholding Overclocking Default clock speed 700Hz
Overclocked to frequencies of:
800Hz
900Hz
950Hz
1000Hz
Dynamic overclocking reduces risk of damage Dynamic Thresholding: Selecting individual threshold value for a pixel based on the range of intensity values in its local neighborhood.
Deals with strong illumination gradients.
Efficient for output results.
Computationally inefficient. Kalman Filter: New algorithm development procedure.
Has two phases:
Predict: Estimates a new state (position) from previous state information.
Update: Information from current state is used to update the prediction procedure.
High accuracy levels.
Complex implementation.
Computationally intensive. Raspberry Pi Camera: Connected via CSI interface
Offers extremely high data rates
Resolution: 5 megapixels
Capable of 1080p and 720p Turret Implementation: A turret system would offer a greater range of view, including tilt and span:
Use the objects current position within the frame to dictate direction of travel
Connected via the GPIO pins available on the Pi Achievements include, implementing an object tracking algorithm with the Raspberry Pi and gaining an extensive insight into the field of image processing
Learning outcomes taken from this project:
Proficient use and understanding of Python, Linux, Debian and Matlab imaging module
Time and Project management Initiatives
Operation and technological insight into the pros and cons of the Raspberry Pi Still image manipulation
Multiple image tracking
Real time image tracking
Algorithm development
Prioritise efficiency for Python
conversion Image Processing using PIL
Numpy - Image array manipulation
Pygame:
Locking, Tracking and Recover algorithm
Position Predict Algorithm
Full transcript