Measuring lesions on leaves

[atalbot]

Apologies if this has been posted before - I searched but couldn't find a complete answer.

I'm trying to analyse the size of disease lesions on the surface of detached plant leaves (see the photo), but I'm struggling to automatically identify lesions.

So far, I've used the Colour Deconvolution tool (http://www.mecourse.com/landinig/software/cdeconv/cdeconv.html) to separate colours, before setting a threshold on the blue channel which can be analysed as particles. Seeing as healthy leaf tissue does not reflect much blue light, this normally works well, however it sometimes completely overlooks lesions (e.g. the second example photo).

Would anyone know of an alternative method of isolating and measuring these lesions? Any ideas are appreciated.

[Olivier Burri]

Hi Unknown,

A few remarks.

Your images are quite ridden with artifacts from the jpeg compression. If you are taking the images, I would recommend you use a RAW format from your camera.

Why do you remove the background? Are there inhomogeneities in it?


Better than doing an RGB deconvolution, what I have seen works OK if converting your image to CIELab (Image->Type->Lab Stack)
Then set a threshold on the second (a*) channel.

run("Lab Stack");
setSlice(2);
run("Gaussian Blur...", "sigma=2 slice");
setAutoThreshold("Moments dark");
run("Analyze Particles...", "size=200-Infinity exclude clear add slice");

If you run the above code, it looks close to what you want.

Best

Oli

> -----Original Message-----
> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of
> atalbot
> Sent: Monday, July 27, 2015 14:43
> To: [hidden email]
> Subject: Measuring lesions on leaves
>
> Apologies if this has been posted before - I searched but couldn't find a
> complete answer.
>
> I'm trying to analyse the size of disease lesions on the surface of detached
> plant leaves (see the photo), but I'm struggling to automatically identify
> lesions.
>
> So far, I've used the Colour Deconvolution tool
> (http://www.mecourse.com/landinig/software/cdeconv/cdeconv.html) to
> separate colours, before setting a threshold on the blue channel which can
> be analysed as particles. Seeing as healthy leaf tissue does not reflect much
> blue light, this normally works well, however it sometimes completely
> overlooks lesions (e.g. the second example photo).
>
> Would anyone know of an alternative method of isolating and measuring
> these lesions? Any ideas are appreciated.
>
> <http://imagej.1557.x6.nabble.com/file/n5013754/LeafLesionExample.jpg>
> <http://imagej.1557.x6.nabble.com/file/n5013754/LeafLesionExample-
> Blue.jpg>
>
>
>
> --
> View this message in context: http://imagej.1557.x6.nabble.com/Measuring-
> lesions-on-leaves-tp5013754.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html