Where is my car?

One of my first image processing based project. It finds cars on a Google Maps view, based on their color.

When I started this project, I had no experience in image processing, so I tried what came to my mind. Looking backward, the sources are a real mess but I acquired my first image processing skills with it.

Here a the basic steps of the program:

  1. Delete the unnecessary colors from the image
  2. Classify the image in 10 colors with the K-means algorithm
  3. Delete huge objects that can't obviously be cars
  4. Fill tiny holes
  5. Classify again
  6. Remove odd shapes
  7. Fill tiny holes again
  8. Restore original colors
  9. Delete pixels with low saturation
  10. Classify again
  11. Remove little objects
  12. Remove odd shapes

Later, I decided to bind it with a Cocoa GUI.

Technologies

The CImg Library: a simple but powerful image processing library
Cocoa: the framework behind every Mac application

Screenshots

Download

Where is my car? 1.1

Version 1.1 (Command line)

You need ImageMagick to build the program.

The Cocoa GUI is not available yet, but you can contact me if you want to give it a try.