Original TIFF image: download
Hi @Kamran, and welcome to the forum!
This forum is all about quantitative image analysis, but not everyone here is familiar with the kind of images you posted.
There’s a rule of thumb that when writing a forum question, you should spend at least the amount of time that you’d expect others to spend with answering it. So, here are a few additional questions to get you started:
- What have you tried already? Did you encounter any specific difficulties?
- How do you (or literature in your field) define pore size?
Also have a look at other topics about pore size here on the forum.
I have tried on ImageJ software but was unable to get the pore size of the image, i found the void fraction that is 31% but i need to find the pore size and its distribution. And I mean from the pore size : the space or volume of the coating that is covered with the solid particles or agglomerates, In this image silica nanoparticles are coated on glass substrate with introduction of porasity to helps in reducing the Refractive index of the silica and also it helps in self cleaning ability.
please read carefully what Jan tried to explain.
I think we are still not able to understand what is what in your image.
Here is an enlarged excerpt of your image:
Please indicate what you consider being a pore and what is what you call void.
I am guessing you are asking this question not because you ran into a problem trying to extract the value but because you just don’t know how. For pore size, I use the Local Thickness plugin which is commonly added to ImageJ via the BoneJ plugin. If you are using Fiji instead of vanilla ImageJ, I believe the plugin is already installed. Note that the results are provided as radii in pixels, not scaled units.
That is awesome @Andrew_Shum. The local thickness result looks very interesting. How do you use the histogram data to determine other properties ?
That’s great @yempski. I really appreciate it. Thanks alot, now how can we find the pore size? I mean the size of blackish part, or porous part?
Hi @Kamran, the histogram has a large peak at zero, so my guess is this represents black pixels. You could take the ratio of this to the total count to determine a percentage. However, I really doubt this is an accurate measure of porosity. Wait for some feedback from @Andrew_Shum
You still didn’t answer the question posed in my earlier post. As long as we don’t know how you define pores, we cannot help. Please refer to the before posted enlarged portion of your image and indicate what you consider being a pore.
Thanks alot for asking question and helping. The white spots shows the nano-particles, while the blackish area shows pores means empty spaces, I want to calculate average size of these pores, or size range in nano-meters. As the nano-particle size is about 10-25 nm, what is pore size or diameter.
The white spots shows the nano-particles, while the blackish area shows pores means empty spaces
Please tell us how you define black and white!
Here is the gray-value histogram of the excerpt of your image:
It shows the number of pixels having a certain gray-value with black to white from left to right. As you can see, the image shows a continuum of gray-values with negligible amounts of pure black and white. Your image shows dark and light “clouds” but they are not black and white!
So what do you call black and white?
Furthermore and related, I can’t find any boundary between what you may call nano-particles and pores.
As long as you can’t provide a formal, (mathematical) definition for nano-particles and pores, there is no way for a machine to do what you want.
Please draw boundaries of nano-particles and pores onto the image excerpt that I’ve posted earlier and post it here.
Hey guys. Sorry for being away from the site for so long.
First, I would like to correct myself. I previously said that the Local Thickness plugin gives you the radius. That is incorrect, it gives you the diameter.
I will correct that in my prior post.
At least from what is here, it isn’t clear that this issue has actually been resolved fully. Herbie is correct in that you will need to clearly define the threshold between pores and not-pores. That is a critical input for the Local Thickness plugin. If you are using the plugin directly, it will give you the opportunity to select the threshold; this means you can operate on gray-scale images. If you are using the plugin through BoneJ, you must segment the gray-scale first. If your segmentation is not based simply on gray-scale value, you will need to segment your image first regardless of which plugin you are using.
To address the question that was posed to me, you can just click the
list button to get a list of all the bin starts and the counts in those bins. If I recall correctly, the lower bounds of the bins are inclusive and the upper bounds are exclusive. An exception is made for the lowest and highest bins which both contain counts for values outside the specified range. I typically set the bounds and number of bins so that the bin starts increment by 0.5 px.
I forgot there is a time limit on editing replies.
I respectfully urge all concerned with this topic to do some research on Fractals, this is what you are looking at. The issue is quite complex and needs to be understood to be analyzed as well as you are attempting.
Why are you bumping this old topic? I respectfully think that this question has nothing to do with fractals and more to do with a misunderstanding of the secondary electron escape depth in SEM.
Just stumbled across it. Thought it was a shame to be left unfinished.
You might be correct about the SEM, got any info on it?
Bob didn’t bump the thread, I did. I realized that neither I nor anyone else ever replied to the question that was posed to me. Furthermore, the there didn’t seem to be a clear resolution.
I have actually checked it out and will send some interesting results shortly (probably tomorrow, its getting late)
Attached you will find a stacked image of a small portion of the original image to give you an idea of what you’re looking at and why it is so important to define to yourself, what you want to obtain as results.
1stSample_Stack_Zcorr.tif smallest structure eigenvalues.zip (8.9 MB)
The image was obtained by slicing the image at each level of the histogram and stacking them from brightest intensity to lowest, and gives you insight into the point that there really is not a clear pattern which could define any specific size. Just use it in the 3D Viewer.
When analyzing an image this way it takes a lot of time and eats a lot of computer space. For instance, Herbs contribution starts at 118Kb size, steps up to 1Mb,then to 64Mb up to over 180Mb just to get to the point that this stack is presented. This stack is from a different place than Herbs but you get the point.
If anyone is still interested I will return shortly with more details.
I decided to go ahead and use what would commonly be considered a ‘pore’ and performed a small analysis on it. Use it as an overlay to the original image and see if you agree with the assumption.
Results.csv (153.3 KB)
I hope it helps someone.
Good day Bob,
could you please explain in formal mathematical terms
what would commonly be considered a ‘pore’