https://github.com/dilawar/object_tracking

https://github.com/dilawar/object_tracking

Last synced: about 23 hours ago
JSON representation

• Host: GitHub
• URL: https://github.com/dilawar/object_tracking
• Owner: dilawar
• Created: 2016-10-15T05:55:36.000Z (about 8 years ago)
• Default Branch: master
• Last Pushed: 2018-05-09T05:20:00.000Z (over 6 years ago)
• Last Synced: 2024-11-09T12:49:36.158Z (about 2 months ago)
• Language: Python
• Size: 61.5 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Object Tracking 

This is an experimental project designed to track a moving animal in an arena.

This may not work out of the box in your case but I'd be happy to help.

# Demo 

See a demo [here](https://youtu.be/Eng82w9g9-w) This demo was created with version v0.0.3 of this script.

# Dependencies 

- python-opencv

- python-numpy 

- gnuplot 

# How to use 

When you run the command `python track.py -f /path/to/recorded_video`, you will be presented the 

first frame to locate the animal. You can drop as many frame as you like by pressing `n`. Once 

you see the animal in frame, use the mouse to draw a rectangle around the animal: press left mouse button

and drag it to another corner then release the button. Press `q` to start tracking.

During tracking, you can click on the animal to fix the location. A white cirlce 

represents the approximate location of animal (see the demo). 

When all frames are done, check the directory of video file, there must be a data file in csv format. First 

column represent the time, second the column index and third the row index of the animal.

# Help

Drop and email to [email protected] . Though you should raise an issue on this repo. Make sure you 

attack a fragment of video for me to test.