Hello.

You may check out the following video for better understanding. It was quite helpful for my project. Hope you find this useful as well.

https://youtu.be/6dGAFawBDMU

Hi Hanaa,

I tested with a calibrated image and it appears that the outside edges are the place to measure from. Although this approach is not completely accurate if the pixel-metric conversion/ scale is not a whole number because ImageJ appears to place the scale bars to cover full pixels. That means that the actual number of pixels the scale occupies can be rounded down.

Perhaps a better way is to obtain the image resolution from the image meta-data somewhere (in the image file, research notes, or acquisition instrument) "if" you need that level of accuracy.

If you have a scale of 1:1, and put a scale bar of size 10 on the image, the scale bar will occupy 10 pixels. However, if the ratio is not 1:1 then a very accurate calculation can not be made from the scale bar.

I checked this on one of my X-ray Ct images at 3.6 um resolution. A bar of 10 um occupies only 2 pixels, so it appears that ImageJ rounds down to the nearest whole pixel. When i choose a bar of 11 um, the bar occupies 3 pixels. So, these scale bars are perhaps more for display than to calculate highly accurate calibrations.

Are the images stored in .tif format? Modern SEM machines can store the metadata in that format and you can load it into Gwyddion (the open source tool that's very popular with AFM people) for further analysis, Gwyddion will then automatically generate new scales.

If your SEM doesn't store as .tif by default, I would recommend to change that, anyway, since most journals prefer microscopy data to be submitted in that format.

Thanks for your answers,

I tried to measure from the outside edges and it gave me the most accurate results when comparing the measurements from images with different magnifications for the same sample.