I would advise the Fiji (just ImageJ - open source http://fiji.sc)

1) Read the Image

2) Threshold the Image

3) Analyze Particles

which parameters do you want to quantify? number of mines, area damaged...

Maybe you can attach an example image?

Anyway, Fiji is a good option to begin with

Thanks for the answers! We have not started the bioassays yet, so we do not have any biological material available to show. A priori we are interested in measuring the length of mines. Hope Fiji works!

This is a bit difficult to answer without more knowledge. If you can observe the damage visually, then surely you could translate that into an image-processing algorithm. My recommendation would be then to use Matlab as it is very quick to prototype a solution.

For this you need image inpainting technique that uses the surrounding texture information to fill the region of interest. there are some image inpainting matlab submission available on mathworks site.

Here you can find an overview of existing image analysis tools for botanical applications: http://www.plant-image-analysis.org/

I think some of them already support scripting for batch processing a lot of leaf images automatically. If you want to develop your own image processing tool, I can highly recommend MATLAB.

You might want to look into CellProfiler. It is a free open source user friendly software designed with life science applications and high throughout put in mind. You can design your own pipeline/algorithm in a GUI, run you images in batch and have the result exported to a spreadsheet or database.

http://www.cellprofiler.org/

Thanks for your advice! http://www.plant-image-analysis.org/ looks really cool!

You may consider some leaf staining techniques that would increase your contrast between the mined area and the green parts.

I think you have to determine the plant of concern. because there different shapes of leaves. Taking this in consideration will help to determine the mined area.