How to remove noise from leaf image

Hi,

I'm asking this questions after spending a lot of days researching without getting an solution for my problem (the good thing is that I've learned a lot).

Well, I have a dataset of leaf images and I will compare their texture to check what's their level of a certain disease. Before working on the code to make comparisons, I've been working on leaf segmentation to remove background (or just make it totally black). I got a binary image and the code was working for some of them, but then I got a bunch of images that have something like a noise in the background and which I'm not able to remove.

I'm pretty new to Matlab, so I'm not sure if this kind of noise is a speckle, salt and pepper or anything else. I just see it when I transform the image to a binary one.

I've attached the original image and the binary image (where you can see the noise in a better way). Most of that it's on top, at left.

Does anyone know how to remove this type of noise from images? I've used medfilter and colfilt, but the results are not good.

Thank you in advance.

Answer 1

[Image Analyst]

You didn't segment the leaf properly. You need to do color segmentation, not whatever you did. See attached demo, below this picture it makes:

[Queila Martins]

Hi,

First of all, thank you so much for your answer, code and images!

In fact I was only transforming the image to binary, so I thought this would work properly (but it didn't). I'm going to study your code and run it to my images dataset. I tried this one with a leaf in a white background and for this one the result is a Leaf Mask Image that's completely white. So, I was wondering if for white background I must try a different approach. What do you think?

Thank you very much for your help!

[parul mittal]

your code file isn't opening . can u plz reload it here.thnx

[Prajwala TM]

Your code gives the following error: Undefined function 'bwareafilt' for input arguments of type 'double'. Error in test (line 103) leafMask = bwareafilt(leafMask, 1); Please help.