Skip to content

metostom/image-classification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
 
 
 
 
 
 
 
 

Repository files navigation

image-classification

Messing Around with some basic Image Segmentation and Classification Using Superpixel Segmentation and K-means classification. Just a fun side project to try out some image classification techniques in an afternoon.

Summary

I downloaded a Eastern Washington RGB GeoTiff and used several superpixel algorithms to segment the landscape. Then basics statistics (mean, min, max, variance, skew...) were computed for each of the red, green, and blue bands within the segements. K-means classification was then used to define groupings within the segments and colored accordingly.

  • Numpy
  • Gdal
  • sci-kit learn

First Test Results

Original Download

Original

Classified Image

Classified

Thoughts

This simple method appears to correctly classify some areas but really does not seem to capture many finer differences in the landscape. Would be interested to apply this to a larger image with more variable landcover. Would need devote more time to accurately segment but not bad for an fun afternoon!

About

Basic Image Segmentation and Classification Using Superpixel Segmentation and K-means classification.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors