Uneven brightness causes trouble in thresholding

Hi,

I am new to ImageJ but didn’t manage to find solutions to this kind of issue:
I am looking at defects in Ti6Al4V and took a lot of OM images. However, the brightness seem to be uneven in themimage
and resulting in this when I converted it to binaryimage
Therefore, I was not able to use the “analyse particles” function to count and measure them.

is there any method to get rid of this issue? Please excuse me if this is already answered.

Hi @seve,

and welcome to the forum.

in such a case you best do a flat-field correction.
Optimally, you take an image without a sample (but with the base your sample normally is on, e.g. a slide with embedding medium and cover slip). I am here assuming that this image comes from a microscope (not sure what you meant with OM image (optical microscopy ?)). Since the uneven lighting comes most likely from your microscopic setup (could potentially be solved by using Köhler illumination during setup (Microscope Alignment for Köhler Illumination | Nikon’s MicroscopyU) or (ZEISS Microscopy Online Campus | Microscopy Basics | Kohler Illumination).

But to solve your issue now, you can install (it is a Fiji update site) the BioVoxxel Toolbox - ImageJ and use the Pseudo-Flat-Field correction.
This is the result after a pseudo flat-field correction using a radius of 60:

A “quick” thresholding shows that the correction will help in easier/improved extraction:

2 Likes

Thanks very much @biovoxxel, problem solved!