find the leaf area

[SadafSafder23]

Hi,

I am working on phenotyping and calculate the leaf area, leaf angle and stem length now I have need to calculate the leaf area using Plantcv, can anyone here to help me what function is used here to calculate each leaf area in a plant. here's the some images of the plant.

Kind Regards,
Sadaf

[nfahlgren]

Hi @SadafSafder23, for the grass plant, following the general steps of this tutorial would get you the measurements you are aiming to do: https://plantcv.readthedocs.io/en/stable/tutorials/sorghum_morphology_tutorial/

The other plant is more complicated for a few reasons. The tools used in the tutorial above work well on thin grass leaves, but skeletonization (as we have it implemented) will not work as intended on the oval-shaped leaves in the first image because you will get a branch segment corresponding to the long axis and another mostly perpendicular corresponding to the short axis. The other issue is that the leaves overlap, which will further break the assumptions of the morphology module, as currently built.

Because the leaves in the first image are nicely spread out to either side of the stem, you may have good luck with watershed segmentation: https://plantcv.readthedocs.io/en/stable/tutorials/watershed_segmentation_tutorial/

[SadafSafder23]

Thank you @nfahlgren for your response, I'm sorry, I'm away that's why could not respond.
I have already follow the step of mentioned tutorials (https://plantcv.readthedocs.io/en/stable/tutorials/sorghum_morphology_tutorial/)
However, I could not understand how to define the leaf area, the formula of leaf area is lenght * width, how we define the width of the sorghum leaf? an other question can we define the length of the stem in centimeter?

in the image, can we change the path length into centimeter?

[HaleySchuhl]

To elaborate on the leaf area algorithm a bit more, the function pcv.morphology.fill_segments is not relying on width or height at all to determine area. Instead it uses the discrete segments as markers for watershed segmentation.

For the stem length, I recommend using the pcv.morphology.analyze_stem which will store the stem height, stem length (describes width usually), and stem angle.

To convert from pixels to centimeter you must have a size marker in your images that have a real, known area/length. Then you can scale the units with simple division once you have your data table.

[SadafSafder23]

thank you for your response @HaleySchuhl, I could not understand the size marker, how can i add in the image?

did you mean like this image, but in this image we take image then scan it by setting a proper scale in scanner rather my work is related to real image that take from RGB image, again thank you

[HaleySchuhl]

It's difficult to use the pcv.report_size_marker function on rulers, especially in this case since the color of the ruler is so similar to the background of your image. The best size markers are usually something like a sticker, a coin, or even a poker chip where you can have a known diameter and area which can be used in downstream analysis to scale pixels in an image to cm-squared.

If I were trying to analyze this specific image, I would measure the ruler using a rectangular ROI to capture the shape and use pcv.analyze_object to measure the ruler. In excel or R after data collection, you can use the actual area (in cm squared) and width (in cms) to scale each 2D and 1D measurements of the plant respectively.

[nfahlgren]

I think there's some confusion here I want to clear up. When Haley mentioned "markers" she wasn't referring to report_size_marker, she means that the pcv.morphology.fill_segments function uses skeleton segments as "markers" for the watershed algorithm to flood fill the leaf and stem segments. The result of this is we count the number of pixels associated with each leaf and stem segment to calculate the area of each

[SadafSafder23]

thank you @HaleySchuhl and @nfahlgren for such a good explanation, it's really helpful