Colored vegetation image analyze with ImageJ

I want to analyze a color image from GGearth and to extract 3 informations: soils (=pathway + reddish color -burnt area-); lawns (clear green) ; forest(dark green).
I tried to binarize and threshold, then to use particule analyzes or measurements.!

I get the following problems:

  1. The background is colored (white or black)
  2. I cannot get 3 categories on the binarized image, but only balck and white
    Somebody gets an idea?

original image with center and corners transparent (does not appear on the screen, but the png images gets transparent center - pool- and corners -outside area of interest-)

ImageJ result with binarize + threshold + measurement

@Catherine_Reeb Perhaps it’s worth giving Trainable Weka Segmentation a try. You’d be able to define classes for each element you are interested in and produce a labeled image to then do your measurements, etc.

1 Like

Thank you, that is exacty what I needed

1 Like